Exposure: Range or Fixed? Which is better?

[Ri22o]

Which exposure method is better to use? A range or a fixed value?

I have seen a lot of comments in regards to exposure where the response was a fixed number and I am questioning whether I should adjust mine or not.

Most, if not all, of my cameras are running around 3.33ms and 20-25 gain during the day. Night settings are around 6ms and 35 gain with all of the IR I have around the house.

Would I be better off setting any of these (mostly all 5442s) to a fixed exposure? I changed my mini-PTZ to a fixed exposure last night and I think it made a difference. (Although I still need to mess with it more.)

 

[bigredfish]

I typically run a range 0-2 or 0-4ms during the day on most of my cameras. For changes in light like dark days with storms vs bright sunshine seems to help.
At night I mostly run fixed 1/120 or faster if I have enough light/IR.
LPR cams always fixed at night

All that said, I do have a few that I run a range at night with no apparent problems

 

[wittaj]

In general, anytime you can force the camera to not have to adjust anything, the better the image as the camera isn't adjusting to the light being presented to it. You can really see that in particular on LPR cams in the day with a shutter and gain range too large - it will pixelate as the vehicles goes thru the tight field of view as it tries to adjust. One of the reasons why we suggest fixed shutters for LPR.

And if you do give it a range, have the difference be as small as possible.

But some field of views are better than others for that.

Depending on the angle and field of view, setting a fixed shutter and gain range tight or the same can be problematic at dawn and dusk as an example.

Say you run a camera at night at fixed shutter of 1/120 and gain 50-55. Works great at night. But before it switches to day profile and it gets lighter out, the 1/120 shutter may be too slow and wash it out.

So in areas that I have overlapping coverage, I try to run one camera with fixed parameters and the other with the range.

And of course the camera has an affect as well. Some cameras do better with ranges than others.

 

[bigredfish]

This 5442 turret runs 4-7ms range at night, 0-50 Gain

View attachment Home_ch4_20231231210102_20231231210113.mp4

 

[Ri22o]

I typically run a range 0-2 or 0-4ms during the day on most of my cameras. For changes in light like dark days with storms vs bright sunshine seems to help.
At night I mostly run fixed 1/120 or faster if I have enough light/IR.
LPR cams always fixed at night

All that said, I do have a few that I run a range at night with no apparent problems

Ok, basically similar to what I have going on.

In general, anytime you can force the camera to not have to adjust anything, the better the image as the camera isn't adjusting to the light being presented to it. You can really see that in particular on LPR cams in the day with a shutter and gain range too large - it will pixelate as the vehicles goes thru the tight field of view as it tries to adjust. One of the reasons why we suggest fixed shutters for LPR.

And if you do give it a range, have the difference be as small as possible.

But some field of views are better than others for that.

Depending on the angle and field of view, setting a fixed shutter and gain range tight or the same can be problematic at dawn and dusk as an example.

Say you run a camera at night at fixed shutter of 1/120 and gain 50-55. Works great at night. But before it switches to day profile and it gets lighter out, the 1/120 shutter may be too slow and wash it out.

So in areas that I have overlapping coverage, I try to run one camera with fixed parameters and the other with the range.

And of course the camera has an affect as well. Some cameras do better with ranges than others.

That all makes sense, and I did consider the concern running fixed during the dawn and dusk times, much the same as with my LPRs.

It looks like I don't really have anything to change, but may play with them some for the experience.

 

[OICU2]

What would be a good custom shutter and gain range for daytime LPR? Right now I am on manual using 1/2000 but do get some pixelating.

 

[wittaj]

What would be a good custom shutter and gain range for daytime LPR? Right now I am on manual using 1/2000 but do get some pixelating.

Depending on sun and direction of LPR, 1/2000 or 1/4000 should be good for daytime.

The pixelation is probably more a result of the gain. 0-50 will cause it.

Again depending on your angle with the sun will dictate the number so try like 20-25 or 30-30 and see if it improves.

 

[OICU2]

Depending on sun and direction of LPR, 1/2000 or 1/4000 should be good for daytime.

The pixelation is probably more a result of the gain. 0-50 will cause it.

Again depending on your angle with the sun will dictate the number so try like 20-25 or 30-30 and see if it improves.

One of mine is almost directly due to west inline with the sun most of the day. It looks fairly washed out most of the time.

 

[wittaj]

One of mine is almost directly due to west inline with the sun most of the day. It looks fairly washed out most of the time.

Then try 1/4000 or a little faster and a lower gain.

 

[OICU2]

What about 3D NR? On or off? (for day LPR) I have it on for day, set to 20 for Advanced 3D and 50 for Advanced 2D.

 

[bigredfish]

I think the Gain range during the day is the biggest factor in pixelization.

I'm back to running 1/2000, 0-50 Gain during the day

20 DNR day for me

 

[wittaj]

Probably needs to be on as these images can pixelate quick with it off, but you can run a low number. I would do 20 for 2D and 3D.

In my experiences, the 2D and 3D need to be the same for most field of views. In theory 2D is for static parts of the image and 3D is for the moving objects, but you can get some halo effects if they spread is too large in some views.

 

[bigredfish]

I just played with it. Gonna try range 0-10 on Gain to see if that helps the pixelization

 

[OICU2]

Sorry for the thread hijack. Here's one of mine just now, still seems a little washed:

 

[bigredfish]

Probably just sun angle. I have a tough one too. I bumped contrast and played with gamma a bit. Between sun angle and tree shadows, I envy folks with nice clean north/south street shots

 

[bigredfish]

As weird as it may sound, my nighttime caps are perfect and consistent. I have way more trouble with daytime as you’re seeing.

 

[OICU2]

My nights are hit and miss. Zero light, zoomed to the max, and night focus/zoom/etc set via BI http presets. Vehicles can be at different spots within the road and the focus might not be where they happen to be.

 

[bigredfish]

You may just not have enough IR on target?

I fought it a lot in the beginning, but ever since I went to using the 'Schedule" profile and forcing it to change to Night/B&W/IR 45 min before sunset, I just dont have any problems. The camera has enough light to get a good focus and sticks.

No plate light


Normal with light


Motorcycle

 

[OICU2]

Quite possibly in addition to shitty focus. I have manual focus set to a specific point, [Edit: I have it to semiauto but BI calls a preset with specific focus/zoom point at night] but I've never actually had a vehicle stop there so I could dial it in. Maybe I need to change it to semiauto or auto? Here's two of mine at night, one around midnight and the other early AM. I do use Schedule as well similar to yours.

 

[bigredfish]

I use Semi Auto